UnifiedSKG: Unifying and Multi-Tasking Structured Knowledge Grounding with Text-to-Text Language Models

Open paper Other version
Benchmark Model Rank Results
semantic-parsing-on-wikitablequestionsT5-3b(UnifiedSKG)#22Accuracy (Test): 49.29Accuracy (Dev): 50.65
table-based-fact-verification-on-tabfactT5-3b(UnifiedSKG)#7Test: 83.68Val: 83.97
table-to-text-generation-on-fetaqaUnifiedSKG (T5-3B)#2BLEU: 33.44
table-to-text-generation-on-fetaqaUnifiedSKG (T5-large)#3BLEU: 32.45
table-to-text-generation-on-fetaqaUnifiedSKG (T5-base)#5BLEU: 29.91
task-oriented-dialogue-systems-on-kvretT5-3b(UnifiedSKG)#1Entity F1: 70.07

Rank counts only results with a code link.